Aggregation of criteria weights for multi-person decision making with equal or different credibility

Abstract

Multi-criteria (MC) problems involve making decision over alternatives that are characterized by several criteria. These criteria represent basis of evaluation in MC evaluation models or goal aspiration in MC optimization models. In most of MC models, criteria weights must be predetermined before the problem can be solved. These weights are interpreted differently but mostly as relative importance of criteria. There are many weighting methods available, but are generally categorized as subjective or objective methods. The subjective methods involve evaluator(s) to evaluate the relative importance of the criteria. Even though multi-person may involve in evaluating the criteria, the final weights must be represented as only one set of weights. Many aggregation methods have been proposed to compose the evaluations. However, these evaluators may have different degree of credibility since they may come from different background or may have different degree of superiority. The aim of this paper is to propose a different concept of weights that would represent the degree of credibility of the evaluators. Furthermore, several aggregation approaches are suggested on how to include these ‘new’ weights in order to produce new criteria weights that also take the credibility of the evaluators into considerations. A numerical example is used to show how these weights of credibility can be used to solve a MC problem in particular to determine the criteria relative importance. This new concept of weight signifies a different insight to the domain of MC decision making (MCDM).